The overall goal of this project is to develop a new collimation technique for single photon emission computed tomography (SPECT) in nuclear medicine. This new collimator merges the sensitivity and resolution properties of pinhole collimation in the transaxial direction with the complete-sampling properties of parallel-beam (PB) and fan-beam (FB). SPECT is a commonly used imaging technique that uses radiolabeled tracers for diagnosing disease states. Collimation is used in SPECT in conjunction with a position-sensitive detector (i.e., a gamma camera). Thus, SPECT measures a set of line integrals, which are inverted through reconstruction software into a 3D distribution of the tracer's concentration. PB and FB, which magnifies in the transaxial direction, are commonly used clinical collimators. Pinhole and cone-beam (CB) collimators are also used in research and clinical scenarios. CB's sensitivity increases near the focal point, analogously to FB's, yet CB's and FB's resolutions are better near the detector. Pinhole collimation is different in that sensitivity and resolution are both better near the focal point (i.e., the aperture), whereas CB's and FB's best resolutions occur where their sensitivities are worst (i.e., near the detector). However, a problem with both pinhole and CB is that they do not yield complete data from a circular orbit, leading to axial blurring artifacts. These artifacts do not occur for PB and FB, which yield complete data. The collimation proposed in this project combines the favorable transaxial sensitivity and resolution properties of a pinhole with the complete-sampling properties of FB. The collimation is 2D in that the collimator may be translated axially yet produce the same projection image; this property is beneficial for image reconstruction because slices are independent of each other, except for detector resolution. This contrasts sharply with pinhole and CB which have large mixing of axial slices. In this project, a prototype 2D-pinhole collimator will be designed and fabricated. The imaging properties of sensitivity and resolution will be calculated analytically and numerically. These calculations will be validated through experimental point-source measurements. The sensitivity and resolution will be incorporated into an iterative reconstruction algorithm that is 3D, but considers only nearby slices since there is limited axial overlap, which will aid reconstruction speed. Reconstruction of experimental phantoms will be used to test the system. In the later years of the project, a set of "production" collimators, each of which is likely to incorporate multiple slits for improved sensitivity, will be fabricated based on the knowledge gained from the prototype. This set will be evaluated for potential clinical impact using anthropomorphic phantoms. The new collimator is expected to be advantageous when the object diameter is mid-size (-11-27 cm). Thus, clinical and research scans that are likely to benefit from this new technique include brain, breast, limb, as well as some pediatric and animal applications. [unreadable] [unreadable] [unreadable]